Data Sheets and MSDSs are intended for different purposes and are usually written by different departments. It is not uncommon for the HSE folks who write the MSDSs to use the most conservative data, such as the internal production specifications, while the marketing folks usually use average (typical) data, rounded in the product's favor.
When I wrote Data Sheets, I calculated the average results from actual production data and rounded to the nearest appropriate increment. For example, in the case of pour points the ASTM D 97 procedure states
"Pour points are expressed in integers that are positive or negative multiples of 3°C". Therefore, if the average of pour point data for a given product was say -52.3°C, it would not be inappropriate to report the next lowest multiple of 3°C or -54°C (at least not inappropriate for a marketer
).
When Our HSE people wrote the MSDSs they typically used internal production specification data to be safe, and this was always significantly more conservative than the average data.
When I wrote internal production specifications, I set these around the
range of the production data rather than the
average. So, using the same example above, the range of pour points from production batches that gave the average of -52.3°C might have been -45°C to -57°C, and in such a case I might have used -45°C for the production specifications. This figure would then be picked up by the HSE folks and used on the MSDS. Hence the difference between the Data Sheet value of -54°C and the MSDS value of -45°C in this example would be 9°C.
This may seem like a wide range, but keep in mind that the reproducibility of the ASTM D 97 procedure is +- 3°C. Add to this the variability of the actual batches, let's say another +- 3°C, and the range of data could be +- 6°C. So, if a batch of product had an actual pour point of say -51°C, a technician can obtain a valid -54°C on his analysis. If the next batch has an actual pour point of -45°C, another technician can obtain a valid -42°C on his analysis. This gives a range of +- 12°C. In reality, lab and batch data tends to be a bit more consistent than this, but these ranges are possible.
These kind of discrepancies in testing and reporting also occur for flash points, where the ASTM D 92 procedure lists a reproducibility of +-16 °F! This is why I often caution against trying to predict oil composition or compare products using such highly variable data.
I have no idea how ExxonMobil treats their data, but am just pointing out that there could be many reasons for discrepancies between Data Sheets and MSDSs other than formulation changes.
Tom NJ